The study introduces a deep generative framework called deepAMP to identify potent, broad-spectrum antimicrobial peptides (AMPs) that can overcome the antimicrobial resistance crisis. deepAMP uses a peptide language model to generate and optimize AMPs, aiming to enhance their membrane-disrupting abilities and reduce resistance. The framework consists of three main components: a pre-trained peptide generative model (deepAMP-general), an AMP optimization model (deepAMP-AOM), and a penetratin optimization model (deepAMP-POM). Through a two-round design and cross-optimization-validation, 18 T1-AMP and 11 T2-AMP candidates were identified, with over 90% showing better inhibition than penetratin against both Gram-positive and Gram-negative bacteria. T2-9, one of the candidates, showed the strongest antibacterial activity comparable to FDA-approved antibiotics. The study also demonstrated that three AMPs (T1-2, T1-5, and T2-10) significantly reduced resistance to S. aureus compared to ciprofloxacin and were effective against skin wound infections in a mouse model. The results highlight the potential of deepAMP in accelerating the discovery of effective, broad-spectrum AMPs for treating drug-resistant bacteria.The study introduces a deep generative framework called deepAMP to identify potent, broad-spectrum antimicrobial peptides (AMPs) that can overcome the antimicrobial resistance crisis. deepAMP uses a peptide language model to generate and optimize AMPs, aiming to enhance their membrane-disrupting abilities and reduce resistance. The framework consists of three main components: a pre-trained peptide generative model (deepAMP-general), an AMP optimization model (deepAMP-AOM), and a penetratin optimization model (deepAMP-POM). Through a two-round design and cross-optimization-validation, 18 T1-AMP and 11 T2-AMP candidates were identified, with over 90% showing better inhibition than penetratin against both Gram-positive and Gram-negative bacteria. T2-9, one of the candidates, showed the strongest antibacterial activity comparable to FDA-approved antibiotics. The study also demonstrated that three AMPs (T1-2, T1-5, and T2-10) significantly reduced resistance to S. aureus compared to ciprofloxacin and were effective against skin wound infections in a mouse model. The results highlight the potential of deepAMP in accelerating the discovery of effective, broad-spectrum AMPs for treating drug-resistant bacteria.